Translational profiling of macrophages infected with Leishmania donovani identifies mTOR- and eIF4A-sensitive immune-related transcripts

Summary

The protozoan parasite Leishmania donovani (L. donovani) causes visceral leishmaniasis, a chronic infection which is fatal when untreated. Herein, we investigated whether in addition to altering transcription, L. donovani modulates host mRNA translation to establish a successful infection. Polysome-profiling revealed that one third of protein-coding mRNAs expressed in primary mouse macrophages are differentially translated upon infection with L. donovani promastigotes or amastigotes. Gene ontology analysis identified key biological processes enriched for translationally regulated mRNAs and were predicted to be either activated (e.g. chromatin remodeling and RNA metabolism) or inhibited (e.g. intracellular trafficking and antigen presentation) upon infection. Mechanistic in silico and biochemical analyses showed selective activation mTOR- and eIF4A-dependent mRNA translation, including transcripts encoding central regulators of mRNA turnover and inflammation (i.e. PABPC1, EIF2AK2, and TGF-β). L. donovani survival within macrophages was favored under mTOR inhibition but was dampened by pharmacological blockade of eIF4A. Overall, this study uncovers a vast yet selective reprogramming of the host cell translational landscape early during L. donovani infection, and suggests that some of these changes are involved in host defense mechanisms while others are part of parasite-driven survival strategies. Further in vitro and in vivo investigation will shed light on the contribution of mTOR- and eIF4A-dependent translational programs to the outcome of visceral leishmaniasis.

Keywords:

Messenger RNA – Parasitic diseases – Macrophages – Protozoan infections – Protein translation – Leishmania donovani – Amastigotes – Promastigotes

Introduction

Visceral leishmaniasis (VL) is a vector-borne infection caused by protozoan parasites of the Leishmania donovani (L. donovani) complex. VL is endemic in more than 60 countries and is frequently lethal if untreated [1]. The lack of efficient vaccines and the failure to control emerging parasite resistance reflect the urgent need to design safe and efficient therapeutics targeting host-encoded factors [2]. In mammalian hosts, Leishmania promastigotes preferentially colonize macrophages, where they transform into replicative amastigotes that proliferate within modified phagolysosomes [1]. To establish a successful infection, the parasite dampens antimicrobial responses, alters vesicle trafficking, and subverts immunomodulatory functions and metabolic processes of the host cell [3].

At the molecular level, L. donovani modulates the activity of multiple host cell signaling pathways and transcription factors [3]. Consistently, profiling of mRNA levels in L. donovani-infected macrophages revealed a vast perturbation in host gene expression programs associated with parasite persistence [4–8]. The pioneering data supporting widespread changes in host cell mRNA levels following L. donovani infection were obtained in bone marrow-derived macrophages (BMDMs) using cDNA-microarrays [4]. This study showed that L. donovani axenic amastigotes downregulate expression of genes involved in apoptosis and NF-κB signaling while stimulating those encoding monocyte chemoattractants. Subsequently, DNA-microarray based studies of human and mouse monocyte-derived macrophages infected with L. donovani promastigotes identified increased levels of transcripts related to cell migration and repression of genes encoding MHC class II molecules [5, 6]. More recently, RNA sequencing (RNAseq) of mouse peritoneal macrophages infected with L. donovani showed a strong suppression of genes related to immune activation, signal transduction, phagosome, and endocytosis [8]. Remarkably, combined analysis of RNAseq data from cells and tissues of infected hamsters provided evidence that despite a strong pro-inflammatory signature in the spleen, L. donovani induced a complex gene expression pattern in splenic macrophages characterized by M1- and M2-associated transcripts that skews their responses to IFN-γ, thereby rendering them more susceptible to the infection [7]. Altogether, these studies support an important role of parasite-directed reprogramming of the host transcriptome in the immunopathogenesis of VL. However, discrepancies between transcriptomics and proteomics data of L. donovani-infected macrophages [9] suggest that post-transcriptional and post-translational mechanisms may also modulate the host cell proteome during VL infection.

Among post-transcriptional mechanisms, transcript-selective changes in translational efficiencies enable cells to swiftly remodel their proteomes in response to environmental cues without requiring de novo mRNA synthesis [10–13]. In eukaryotes, translational efficiency is mainly regulated at the initiation step when ribosomes are recruited to the mRNA [14]. This process is facilitated by the eukaryotic translation initiation factor 4F (eIF4F) complex, consisting of eIF4E, the mRNA 5’-m7G-cap-binding subunit; eIF4G, a scaffolding protein; and eIF4A, an RNA helicase [14]. Activation of the mechanistic target of rapamycin (mTOR) complex 1 stimulates formation of the eIF4F complex which promotes translation of mRNAs that are particularly sensitive to changes in eIF4E levels and/or availability. These include those containing a 5’ terminal oligopyrimidine (5’ TOP) motif [15, 16], those with highly structured 5’ UTR sequences, which are largely dependent on the RNA helicase activity of eIF4A for their translation [17], and those with very short 5’UTRs [16, 18]. Many mTOR- and eIF4A-sensitive mRNAs encode proteins related to translation, cell survival, metabolism, proliferation, and growth [16, 18, 19]. Interestingly, a number of innate immune regulators are also under translational control via mTOR- or eIF4A-dependent mechanisms [20, 21], including several pro- and anti-inflammatory mediators in macrophages [11–13, 22]. Thus, key immune cell functions may be hijacked by intracellular pathogens via modulation of mRNA translation. Here we show that infection with promastigotes or amastigotes of L. donovani leads to an early translational reprogramming in macrophages partially depending on mTOR and eIF4A activity which appears to contribute to both parasite persistence and host cell defense.

Results

L. donovani selectively modulates the macrophage translatome

To investigate the impact of L. donovani infection on the host cell translatome (i.e. the pool of efficiently translated mRNAs) at a transcriptome-wide level, BMDMs were incubated with L. donovani promastigotes or amastigotes for 6 h and compared to uninfected cells using polysome-profiling quantified by RNAseq (Fig 1A). Polysome-profiling determines levels of both efficiently translated mRNA and cytoplasmic mRNA. Such data enables identification of bona fide changes in translation efficiencies (i.e. changes in levels of polysome-associated mRNA which are not paralleled by corresponding changes in cytoplasmic mRNA levels) using the anota2seq algorithm [23]. At a false discovery rate (FDR) ≤0.15, anota2seq identified widespread mRNA-selective alterations in translational efficiencies upon infection with either parasite life stage (Fig 1B). From a total of 9,442 host protein-encoding mRNAs detected, 27% showed altered translational efficiency following infection with L. donovani amastigotes (13% increased and 14% reduced) (Fig 1C, left panel; Fig 1D top panels, and S1 Table). Similarly, the translational efficiency of 18% of the host cell transcripts was altered in response to L. donovani promastigotes (9% increased and 9% decreased) (Fig 1C, right panel; Fig 1D, bottom panels, and S1 Table). Consistent with changes in translational efficiencies largely independent of parasite stage, only 1.5% of all mRNAs differed between host cells infected with L. donovani promastigotes as compared to amastigotes (S1 Fig and S1 Table). In addition to detecting differences in translational efficiency predicted to affect protein expression, anota2seq allows for identification of transcripts whose changes in mRNA abundance are buffered at the level of translation such that their polysome-association remains largely unaltered [23]. This is a mode for regulation of gene expression which offsets the relationship between mRNA levels and protein levels to suppress changes in protein levels imposed by altered transcription or mRNA stability [24]. Interestingly, a large subset of transcripts whose abundance changed upon infection with L. donovani amastigotes or promastigotes was buffered at the level of translation (21% out of 1,051 and 29% out of 1,604 mRNAs, respectively) (Fig 1D and S1 Table). In contrast, only a small number of transcripts (71 mRNAs) whose levels differed between L. donovani promastigote- and amastigote-infected BMDMs (S1B and S1C Fig) were translationally buffered. Thus, both life stages of L. donovani induce abundant and largely similar selective changes in translational efficiency of host cell mRNAs that modulate or maintain protein levels. Polysome tracings experiments indicated an increase in translation initiation rates in BMDMs incubated with live parasites, as evidenced by a shift in the proportion of mRNAs found in the monosomal fractions (i.e. containing mRNAs that are not translated or poorly translated) towards the heavy polysome fractions, which contain mRNAs that are being efficiently translated (S2A Fig, left panel and S2B Fig). In contrast, no apparent changes in polysome-to-monosome ratios were detected in BMDM incubated with heat-killed parasites or latex beads (S2A Fig, middle and right panels, respectively, and S2B Fig). These data provide evidence that modulation of host cell mRNA translation initiation is specific to BMDM infection with live L. donovani parasites.

Transcript-selective changes in translation upon L. donovani infection target a variety of macrophage functions

To assess how the host cell phenotype is potentially affected by mRNA-selective perturbation of translation during L. donovani infection, we searched for enrichment of cellular functions defined by Gene Ontology (GO) classifications [25] among proteins encoded by transcripts showing altered translational efficiencies (Fig 2A and S2 Table). Enriched categories, among translationally activated transcripts in BMDMs infected by either parasite life stage, included chromatin remodeling, regulation of mRNA metabolism (i.e. splicing, export from the nucleus, stability and translation), regulation of type I IFN production and protein deubiquitination (Fig 2A, left panel; FDR ≤0.05). Accordingly, mRNAs encoding histone modifying enzymes (e.g. Ash1l, Ep300, Kmt2a, Kmt4c), transcription factors (e.g. Cebpb, Foxo4, Ets2, Elk4), translation initiation factors (e.g. Eif3a/b/c, Eif4g3), ribosomal proteins (e.g. Rpl13a, Rpl38, Rps12), RNA-binding proteins (RBPs) (e.g. Pabpc1, Eif2ak2, Larp1, Pum1), and ubiquitin hydrolases (e.g. Usp25, Fam63b, Usp36) showed increased translational efficiency in L. donovani-infected macrophages compared to uninfected controls (log₂ fold-change ˃1.0, FDR ≤0.15) (Fig 2B, top panels). In contrast, proteins encoded by mRNAs whose translational efficiency was suppressed upon infection by either life stage of L. donovani were enriched for categories related to protein trafficking (i.e. Rab protein signal transduction, vesicle organization, and post-Golgi vesicle-mediated transport), cell metabolism (i.e. mitochondrial membrane organization, mitochondrial respiratory chain complex assembly, fatty acid beta-oxidation, and peroxisomal membrane transport), protein ubiquitination, and tRNA metabolism (Fig 2A, right panel; FDR ≤0.05). Specifically, mRNAs translationally suppressed by promastigotes or amastigotes of L. donovani encode proteins involved in antigen presentation (e.g. Cd74, H2-Q1), intracellular transport (e.g. Rab18, Sec22a, Vamp3, Vps37c), organization of lysosome (e.g. Bloc1s2, Laptm4b), mitochondria (e.g. Cox18, Ndufa8, Timm21, Tomm22), peroxisome (e.g. Paox, Pex2, Pex7) and Golgi apparatus (e.g. Golga7, Tango2), lipid metabolism (e.g. Nr1h2, Pla2g6, Scp2), protein ubiquitination (e.g. Trim68, Ube2g1, Ube2w), and tRNA modifications (e.g. Trmo, Tsen15, Trmt12) (log₂ fold-change < -1.0, FDR ≤0.15) (Fig 2B, bottom panels). Interestingly, several mRNAs encoding innate immune sensors showed activated translation (e.g. Dhx9, Tlr9, and Zbtb20) while others were translationally suppressed (e.g. Clec7a, Mavs, Tlr2).

L. donovani infection enhances mTOR-sensitive mRNA translation in macrophages

We next sought to identify upstream mechanisms underlying observed changes in selective mRNA translation. Activation of Akt and ribosomal protein kinase 1 (S6K1) has been reported in macrophages as early as 30 min post-infection with L. donovani promastigotes [26]. As altered activity of the PI3K/AKT/mTOR pathway modulates mRNA translation in a selective fashion [16, 18, 27], we characterized the kinetics of mTOR activity upon L. donovani infection. To this end, we monitored the phosphorylation status of its downstream targets S6K1 and eIF4E-binding protein 1 (4E-BP1) in BMDMs. Transient phosphorylation of S6K1 at T389 was observed in BMDMs infected by either L. donovani promastigotes or amastigotes (Fig 3A). Accordingly, phosphorylation of S6K1 substrate ribosomal protein S6 (RPS6) at S235, S236, S240, and S244 was augmented during infection (Fig 3A). In addition, early phosphorylation of 4E-BP1 at T37/46 was induced with similar kinetics by both parasite life stages (Fig 3A). To assess whether mTOR-sensitive translation was regulated in L. donovani-infected BMDMs, we focused on 5’ TOP-containing mRNAs, whose translation is highly dependent on mTOR activity and encode ribosomal proteins and translation initiation and elongation factors [16, 18, 27]. Indeed, translation of previously described TOP-mRNAs [27] was selectively activated during infection independent of parasite stage (p < 0.001 for both stages) (Fig 3B and 3C and S3A Table). Thus, PI3K/AKT/mTOR signalling and downstream mTOR-sensitive translation are activated in BMDMs early during infection by L. donovani amastigotes or promastigotes.

Translation of mRNAs encoding RNA-binding proteins PABPC1 and EIF2AK2 is activated during L. donovani infection in an mTOR-dependent fashion

A number of studies point to a central role for RBPs in coordinating macrophage inflammatory responses and anti-microbial activity [28], including poly(A)-binding protein cytoplasmic 1 (PABPC1) [29]and protein kinase dsRNA-activated (PKR, also known as EIF2AK2) [30, 31]. Remarkably, anota2seq detected increased translational efficiency of Pabpc1 and Eif2ak2 mRNAs in BMDMs infected with L. donovani promastigotes or amastigotes (Fig 2B, top-middle panel and S1 Table). Accordingly, expression of both proteins increased upon infection (Fig 4A) without significant changes in mRNA abundance (Fig 4B). Moreover, phosphorylation of EIF2AK2 at T451 was augmented in response to infection with either parasite stage (S2 Fig). Surprisingly, phosphorylation of eIF2α at S51, a main downstream target of EIF2AK2 [32], decreased in L. donovani-infected BMDMs (S3 Fig). The Pabpc1 mRNA was previously shown to contain a TOP-motif [15] and, consistently, treatment with mTOR inhibitors rapamycin or Torin-1 suppressed PABPC1 protein expression during L. donovani infection (Fig 4C) independently of mRNA levels (Fig 4D, top panel). Of note, neither host cell [11] nor extracellular parasite viability (S4 Fig) were affected by mTOR inhibition. EIF2AK2 is induced by TLR stimuli such as LPS and poly (I:C) which also activate mTOR signaling [33, 34]. Moreover, L. donovani activates TLR signaling [35]. Interestingly, L. donovani-induced EIF2AK2 protein expression in BMDMs was reduced in presence of rapamycin or Torin-1 (Fig 4C) independently of Eif2ak2 mRNA levels (Fig 4D, bottom panel). In sum, this set of experiments indicates that L. donovani infection activates translation of Pabpc1 and Eif2ak2 mRNAs to increase PABPC1 and EIF2AK2 protein levels in BMDMs by stimulating mTOR activity.

Translation of eIF4A-sensitive mRNAs is activated upon L. donovani infection

The RNA helicase eIF4A facilitates translation of transcripts harboring long and highly structured 5’ UTR sequences. Some such encoded proteins are involved in tumor immune evasion [36] and progression of viral [37] and protozoan parasitic infections [12], suggesting that eIF4A-dependent translation may contribute to herein observed changes in translational efficiencies (Fig 1). In addition to eIF4F-complex formation, the unwinding activity of eIF4A is enhanced by the translation initiation factor eIF4B [14]. Consistent with eIF4B-dependent modulation of eIF4A activity, levels of phosphorylated and total eIF4B protein were increased in BMDMs infected with L. donovani amastigotes or promastigotes (Fig 5A). To test whether this may contribute to selective regulation of mRNA translation following parasite infection, we assessed translational efficiencies of a compilation of previously described eIF4A-sensitive transcripts [36, 38–40]. Indeed, following infection independently of parasite stage, the translational efficiencies of such mRNAs were elevated as compared to background transcripts (p <0.001) (Fig 5B). From a total of 1198 previously described eIF4A-sensitive mRNAs, 149 were translationally activated whereas 80 were translationally suppressed upon infection with promastigotes or amastigotes of L. donovani (S3B Table). The presence of a 5’ UTR G-quadruplex-forming guanine quartet (CGG)₄ motif is an indirect approach to assess whether transcripts are expected to be more dependent of eIF4A for their translation [41]. Indeed, analysis of Motif Enrichment (AME) revealed a significant enrichment of the (CGG)₄ motif in 5' UTRs of transcripts with highly activated translation (≥4-fold increase in translational efficiency upon infection) as compared to 5’ UTRs from transcripts with unaltered translational efficiency (p = 0.0036) (Fig 5C). TGF-β is a key cytokine implicated in the distinctive immune suppression that follows L. donovani infection [42]. Upon infection with L. donovani amastigotes or promastigotes, anota2seq analysis identified augmented translational efficiency of the tumor growth factor-β 1 (Tgfb1) mRNA (Fig 2B, top-right panel and S1 Table), which is highly dependent of eIF4A for its translation [39]. Accordingly, production of TGF-β increased in BMDMs upon infection (Fig 5D) without changes in Tgfb1 mRNA abundance (Fig 5E). Remarkably, inhibition of eIF4A activity using silvestrol abrogated TGF-β induction in L. donovani-infected BMDMs (Fig 5D) without affecting Tgfb1 mRNA levels (Fig 5E). Of note, no acute toxicity was detected in BMDMs and extracellular parasites exposed to silvestrol (S5 Fig). Thus, these data indicate that, in addition to mTOR-dependent translation, L. donovani infection also bolsters eIF4A-sensitive translation of selected host cell transcripts, including the one encoding the immunomodulatory cytokine TGF-β.

L. donovani survival within macrophages is differentially modulated through mTOR and eIF4A activity

During the course of infectious diseases, translational control acts as a host defense mechanism but can also be exploited by the invading pathogen as a survival strategy [43]. In regard to infections caused by protozoan parasites, augmented mTOR-sensitive translation was associated with parasite persistence during Toxoplasma gondii infection in macrophages [44] whereas eIF4A inhibition suppressed progression of cerebral malaria [12]. These findings in combination with activation of mTOR- and eIF4A-sensitive translation in BMDMs upon parasite infection (Fig 3 and Fig 5, respectively), prompted us to investigate the role of these translational regulators for L. donovani survival within the host cell. To address this, BMDMs were pre-treated with either rapamycin or silvestrol and infected with L. donovani promastigotes. Interestingly, parasite numbers increased in presence of rapamycin at 24 h post-infection (~92% increase compared to DMSO control) (Fig 6A) whereas the opposite effect was observed upon cell exposure to silvestrol (~57% reduction compared to DMSO control) (Fig 6B). These data indicate that mTOR limits L. donovani persistence within the host cell while eIF4A promotes it.

Discussion

Transcriptome-wide analyses of mRNA translation in infected cells and tissues have revealed a profound yet selective perturbation of the host translatome during infections caused by intracellular pathogens, including viruses and the protozoan parasite Toxoplasma gondii [44–46]. L. donovani infection causes widespread changes in host gene expression [4–7, 42]. Yet, differences between the transcriptome and the proteome of infected cells suggested that post-transcriptional mechanisms may also contribute to establishing the post-infection proteome [9]. Herein, using polysome-profiling, we demonstrate early translational reprogramming in macrophages following infection by L. donovani amastigotes and promastigotes. The majority of the changes in the translatome of the host cell were induced in a similar fashion by the two life stages of the parasite, including activation of both mTOR- and eIF4A-sensitive translation programs encoding central regulators of inflammation and mRNA metabolism. Inhibition of mTOR promoted parasite survival within the host cell whereas blockade of eIF4A activity had the opposite effect. These data suggest that some of the changes detected in host mRNA translation during L. donovani infection could be involved in host defense mechanisms while others might be part of parasite-driven survival strategies.

Previous high-throughput studies identified numerous biological processes affected in macrophages during L. donovani infection [5, 6, 8, 9]. Translational profiling of infected macrophages indicates that in addition to altered mRNA levels, L. donovani selectively adjusts the proteome of the host cell to its own benefit by modulating the translational efficiencies of subsets of mRNAs. In line with dysregulated antigen presentation during L. donovani infection [47], transcripts encoding several MHC class I components were translationally suppressed in L. donovani-infected macrophages. In contrast, translation of mRNAs related to chromatin remodeling (i.e. histones and DNA- and histone-modifying enzymes) was augmented in cells infected with L. donovani. This suggests translational control of parasite-directed epigenetic changes known to inhibit innate immune responses of the host cell [48]. Supporting this notion, a proteomics study identified several histones and chromatin remodelling proteins induced in macrophages during L. donovani infection which correlated with greater transcriptional activity [9].

Additional biological processes targeted during VL are related to host mRNA metabolism (i.e. stability, splicing and translation) [9, 49]. Consistently, we identified numerous macrophage transcripts encoding translation initiation factors, splicing factors and RBPs with altered translational efficiencies during L. donovani infection. The mouse genome encodes more than a thousand of RBPs [50], including a subset we identified as translationally activated upon L. donovani infection. Regulation of RBPs is of particular interest during VL as they play a central role during immune responses [28]. Further investigation is required to assess the impact of translational control of host RBP and mRNA metabolism for the outcome of L. donovani infections.

Consistent with reports on early activation of PI3K/Akt and/or mTOR signaling during L. donovani infection in human monocyte-derived macrophages as well as human monocytic-like THP-1 and mouse macrophage RAW 264.7 cell lines [26, 51, 52], we observed increased phosphorylation of mTOR downstream targets in BMDMs as early as 2 h and up to 8 h post-infection. Moreover, our data showing no change in the phosphorylation status of S6K1 and 4E-BP1 between 12 and 24 h post-infection are in line with a study carried out in THP-1 infected with L. donovani for 24 h [53]. Thus, rapid and transient activation of the mTOR pathway appears to be a hallmark of L. donovani infection in mouse and human macrophages. Herein, this event was paralleled by a significant increase in translational efficiencies of a large number of mTOR-sensitive transcripts characterized by the presence of a 5’ TOP motif. In particular, we found that PABPC1, an RBP whose encoding mRNA harbors a TOP-motif [15], is induced during infection in an mTOR-dependent fashion. PABPC1 regulates stability and translation of mRNAs [54]. Notably, PABPC1 is part of an inhibitory translational complex along with the zing finger protein 36 (Zfp36) that prevents overexpression of pro-inflammatory mediators in activated macrophages [29]. Therefore, it is plausible that PABPC1 binds to specific host mRNAs to dampen macrophage inflammatory responses during L. donovani infection.

In addition to PABPC1, our data indicate that EIF2AK2 is upregulated in L. donovani-infected macrophages through mTOR-dependent mechanisms. Given that L. donovani triggers TLR signaling in macrophages [35] and that both mTOR signaling and EIF2AK2 expression are augmented in response to TLR stimulation [34, 55], TLR-dependent mTOR activation might account for increased EIF2AK2 synthesis in L. donovani-infected macrophages. Accumulating evidence suggests that the role of EIF2AK2 during Leishmania spp. infection varies between parasite species. For example, EIF2AK2 activity is induced during L. amazonensis infection and promotes parasite survival within macrophages through induction of an Nrf2-dependent antioxidant response [30, 56]. Conversely, L. major prevents the activation of EIF2AK2 to avoid parasite clearance via EIF2AK2-inducible TNF production [30, 31]. We observed that eIF2α phosphorylation decreases upon infection despite upregulation in EIF2AK2 expression and phosphorylation. This may be explained by the concomitant increase in mTOR activity which can recruit NCK1 to eIF2β and thereby reduce eIF2α phosphorylation to bolster ternary complex formation [57]. Thus, the relative activity of eIF2α kinases and phosphatases may determine eIF2α phosphorylation-status and thereby downstream effects on gene expression and biology. Herein, regardless of the precise mechanism, the increased level of EIF2AK2 does not appear activate the integrated stress response [58].

Our data showing that rapamycin favors parasite survival within macrophages suggest that mTOR- activation is part of a host defense mechanism against L. donovani. These observations are in agreement with a report in which macrophage exposure to rapamycin enhanced intracellular survival of another Leishmania sp. (i.e. L. major) [59]. In contrast, rapamycin treatment has been shown to prevent parasite replication within macrophages infected with various Leishmania spp. [60, 61], including L. donovani [53]. However, this apparent discrepancy might stem from the fact that much higher concentrations of rapamycin (IC₅₀ ≥ 8μM) were used in the studies reporting a leishmanicidal effect for this compound (i.e. at least 400-fold difference).

Early phosphorylation of eIF4B and enrichment of reported eIF4A-sensitive mRNAs among translationally activated transcripts support that eIF4A also affects the translatome of L. donovani-infected macrophages. The activity of eIF4A facilitates translation of transcripts encoding immune modulators (e.g. CXCL10, IRF1, IRF7, iNOS, STAT3, TGF-β, etc.) in various cell types [18, 20, 39, 40, 62] including macrophages [12, 22]. Similarly, we observed that translation of eIF4A-sensitive transcript Tgfb1 and subsequent TGF-β production were upregulated in L. donovani-infected in macrophages. TGF-β is a cytokine associated with immune suppression and resistance to treatment during human and experimental VL [42]. In view of these previous studies and our present findings, it is conceivable that eIF4A-sensitive translation contributes to the immuno-pathogenesis of L. donovani infection. Further characterization of eIF4A translational outputs in L. donovani infected cells and tissues will shed light on this matter.

We observed that silvestrol-mediated inhibition of eIF4A dampened L. donovani survival within macrophages. Interestingly, eIF4A inhibitors have been shown to act as dual-targeting therapeutic agents by interfering with both pathogen and host eIF4A activity during malaria, an infection caused by protozoan parasites of the genus Plasmodium [12]. Of note, a strain of L. donovani resistant to miltefosine had increased levels of eIF4A [63] and high-throughput screenings identified Leishmania eIF4A as a potential druggable target [64–67]. Collectively, our data along with previous studies warrant further investigation on the potential of targeting eIF4A-dependent host and parasite mRNA translation to treat L. donovani infections.

The outcome of VL is defined by a complex network of converging molecular events at the interface between the parasite and the host [1]. Herein, we have uncovered vast reprogrammed translation of protein-coding mRNAs expressed in macrophages upon infection by L. donovani amastigotes or promastigotes. Accordingly, numerous host transcripts critical during infection, including regulators of mRNA metabolism and inflammation, were found to be under translational control. Notably, our data indicate that some of these changes contribute to parasite clearance whereas others favor parasite persistence within the host cell, hinting at the therapeutic potential of perturbing specific host translation programs to control the infection.

Materials and methods

Reagents

Culture media and supplements were purchased from Wisent, Gibco, and Sigma-Aldrich; rapamycin was obtained from LC Laboratories; Torin-1 was provided by Cayman; 10-phenanthroline monohydrate was acquired from Sigma-Aldrich; silvestrol was purchased from Biovision.

Parasites

L. donovani (LV9 strain) amastigotes were isolated from the spleen of infected female Golden Syrian hamsters (Harlan Laboratories) as previously described [68]. L. donovani (LV9 strain) promastigotes were differentiated from freshly isolated amastigotes and were cultured at 26°C in M199 medium supplemented with 10% heat-inactivated FBS, 100 μM hypoxanthine, 5 μM hemin, 3 μM biopterin, 1 μM biotin, 100 U/mL penicillin, and 100 μg/mL streptomycin. Early passage stationary phase promastigotes were used for macrophage infections.

Ethics statement

Housing and experiments were carried out under protocols approved by the Comité Institutionnel de Protection des Animaux (CIPA) of the INRS–Centre Armand-Frappier Santé Biotechnologie (CIPA 1308–04 and 1710–02). These protocols respect procedures on good animal practice provided by the Canadian Council on animal care.

Differentiation of bone marrow-derived macrophages

Bone marrow-derived macrophages (BMDMs) were generated from precursor cells from murine bone marrow, as previously described [44]. Briefly, marrow was extracted from bones of the hind legs, red blood cells were lysed, and progenitor cells were resuspended in BMDM culture medium supplemented with 15% L929 fibroblast-conditioned culture medium (LCCM). Non-adherent cells were collected the following day and were cultured for 7 days in BMDM culture medium supplemented with 30% LCCM with fresh medium replenishment at day 3 of incubation.

Infection of bone marrow-derived macrophages

BMDM cultures were inoculated with L. donovani promastigotes or amastigotes at a multiplicity of infection (MOI) of 10:1, as previously described [69]. Prior to infection, cells were serum-starved for 2 h and treated with inhibitors, when applicable.

Viability assays

Viability of BMDMs and extracellular L. donovani parasites was determined by the resazurin assay as described [11]. Briefly, cells were treated with increasing concentrations of rapamycin (2.5–160 nM), Torin-1 (12.5–800 nM), silvestrol (0.8–100 nM) or an equivalent volume of DMSO (vehicle) for 24 h at 37°C, 5% CO₂. Medium was removed and replaced with fresh culture medium supplemented with 0.025% resazurin. Cultures were incubated for 4 h in presence of the inhibitors or DMSO at 37°C, 5% CO₂. Optical density was measured using a Multiskan GO (Thermo-Fisher) at 600 and 570 nm. Absorbance at 600 nm was subtracted from readings at 570 nm. Experiments were performed in biological replicates (n = 2); each sample was analyzed in a technical triplicate, the average of which was plotted against increasing concentrations of the respective inhibitor.

Quantification of intracellular parasites

BMDM cultures were treated with 20 nM rapamycin, 25 nM silvestrol or an equivalent volume of DMSO (vehicle) for 2 h, then inoculated with L. donovani parasites (MOI 10:1) for 6 and 24 h. Relative quantification of intracellular parasites was performed by RT-qPCR measuring the expression of Leishmania Kmp11 gene as described [70]. Primer sequences are listed in S4 Table. Primer efficiency was calculated using ThermoFisher's online qPCR Efficiency Calculator software (S6 Fig).

Quantitative RT-PCR

Purified RNA (500 ng) was reverse transcribed using the Superscript IV VILO Master Mix (Invitrogen). Quantitative PCR was performed with PowerUp™ SYBR^® Green Master Mix (Applied Biosystems). Relative quantification was calculated using the comparative Ct method (ΔΔCt) [71] and relative expression was normalized to mouse Actb. Experiments were performed in independent biological replicates (n = 3); each sample was analyzed in a technical triplicate, the average of which was plotted against the respective conditions used. Primers were designed using NCBI Primer-BLAST (http://www.ncbi.nlm.nih.gov/tools/primer-blast/) (S4 Table).

Western blot analysis

Following infection and other treatments, total cell lysates were prepared for SDS-PAGE and Western blotting as described [72]. Primary antibodies anti-mTOR (#2983), anti-phospho-S6K1 (T389; #9234), anti-phospho-RPS6 (S235/236; #2211), anti-phospho-RPS6 (S240/244; #5364), anti-phospho-eIF4B (S422, #3591), anti-eIF4B (#3591), anti-phospho-4E-BP1 (T37/46; #2855), anti-4E-BP1 (#9644), anti-phospho-eIF2α (S51; #3398), anti-eIF2α (#2103), anti-PABCP1 (#4992), and β-actin (#3700) were purchased from Cell Signaling Technologies; anti-phospho-PKR (EIF2AK2) (T451; #07–886) was obtained from Millipore; anti-RPS6 (#sc-74459), anti-S6K1 (#sc-230), and anti-PKR (EIF2AK2) (#sc-6282) were acquired from Santa Cruz Biotechnology. Horseradish peroxidase (HRP)-linked goat anti-rabbit and goat anti-mouse IgG secondary antibodies were purchased from Sigma-Aldrich.

RNA fractionation and purification from polysome fractions

Cytosolic lysates of infected and control BMDMs were prepared for RNA fractionation as described [44]. Lysates were layered over 5 to 50% sucrose density gradients and sedimented using a Beckman SW41 rotor at 36,000 rpm (= 221,830.9 × g) for 2 h at 4°C. Gradients were fractionated and collected (30 s, 500 μL/fraction), and the absorbance at 254 nm was recorded continuously using a Brandel BR-188 density gradient fractionation system. Input material (total cytoplasmic mRNA) and efficiently translated mRNA (heavy polysome-associated, >3 ribosomes) were extracted with QIAzol (Qiagen) and purified using RNeasy MinElute Cleanup Kit (Qiagen). Purity and integrity of RNA was assessed using a Bioanalyzer 2100 with a Eukaryote Total RNA Nano chip (Agilent Technologies).

RNAseq data processing

RNAseq libraries were generated using the Smart-seq2 method [73] to enable polysome-profiling of small samples as described previously [74]. Libraries were sequenced by using an Illumina HiSeq2500 instrument with a single-end 51-base sequencing setup for both total cytoplasmic and polysome-associated mRNAs (mRNAs associated with >3 ribosomes) from three independent biological replicates for uninfected and L. donovani promastigote-infected BMDMs, and five independent biological replicates for L. donovani amastigote-infected BMDMs. First, RNAseq reads mapping to the reference genome of the Nepalese BPK282A1 strain of L. donovani were removed (12.7% and 1.4% mappings on average for promastigotes and amastigotes, respectively). The filtered reads were then aligned to the mouse genome mm10. HISAT2 was used for all alignments with default settings [75]. Gene expression was quantified using the RPKMforgenes.py script (http://sandberg.cmb.ki.se/media/data/rnaseq/rpkmforgenes.py) [73] with options -fulltranscript -readcount -onlycoding from which raw per gene RNAseq counts were obtained (version last modified 07.02.2014). Genes that had zero counts in all samples were discarded. Annotation of genes was obtained from RefSeq.

RNAseq data analysis using anota2seq

RNAseq counts were normalized within anota2seq using default TMM-log2 normalization [23]. Significant changes in translation, abundance and buffering were identified by anota2seq [23] using default parameters with the following modifications: minSlopeTranslation = -0.5; maxSlopeTranslation = 1.5; FDR ≤0.15; apvEff >log₂(2.0); deltaPT>log₂(1.25); and deltaP>log₂(1.5). In anota2seq model one, infections were compared to control (i.e. Ld PRO versus control and Ld AMA versus control); in model two, cells infected by different parasite stages were compared together with a contrast to complete the anota2seq model (i.e. Ld PRO vs Ld AMA and Ld PRO versus control). Identifiers for genes which cannot be distinguished based on their high sequence similarity (also reported by RPKMforgenes.py), were excluded from downstream analyses. For further analysis, we used a published list of TOP-containing mRNAs [27] and previously reported translational signatures of eIF4A-sensitive mRNAs [36, 38–40]. The difference in log₂ fold-change of translational efficiency (i.e. apvEff) between the signatures and the background was assessed using Wilcoxon-Mann-Whitney tests.

Gene ontology analyses

Gene ontology analyses were performed using the Panther tool [25] of the Gene Ontology Consortium (http://geneontology.org/) on the union of transcripts activated or inhibited in BMDMs infected by L. donovani amastigotes or promastigotes. Heatmaps of translational efficiencies of transcripts activated or inhibited in BMDMs infected by L. donovani amastigotes or promastigotes were generated using Morpheus.

(https://software.broadinstitute.org/morpheus/index.html, Broad Institute)

RNA sequence motif analyses

Non-redundant RefSeq 5' UTRs were retrieved from genome build mm10 using the UCSC Table Browser (https://genome.ucsc.edu/cgi-bin/hgTables). Analysis of Motif Enrichment (AME) was performed on the RefSeq-annotated 5’ UTR sequences from transcripts with ≥4-fold increases in TE in L. donovani-infected BMDMs (402 5’ UTRs) compared to a control list of 228 randomly selected 5’ UTRs from the set of non-translationally regulated transcripts. The presence of the previously described guanine quartet (CGG)₄ motif [39] in both lists was scored and a one-tailed Fisher’s exact test was performed to determine significance of enrichment.

Statistical analysis

Where applicable, data are presented as the mean ± standard deviation (SD) of the mean. Statistical significance was determined by using one-way ANOVA followed by a Tukey post-hoc test; calculations were performed by using Prism 7 software package (GraphPad). Differences were considered significant when *p < 0.05, ** p < 0.01, *** p < 0.001.

Supporting information

S1 Fig [a]
Differential regulation of host mRNA translation between . promastigotes and amastigotes.

S2 Fig [a]
Polysome tracings of BMDMs infected with live or heat-killed . promastigotes or treated with latex beads.

S3 Fig [tif]
. infection promotes EIF2AK2 but not eIF2α phosphorylation.

S4 Fig [tif]
Measurement of acute toxicity of mTOR inhibitors on extracellular . promastigotes.

S5 Fig [a]
Measurement of acute toxicity of silvestrol on BMDMs and . promastigotes.

S6 Fig [tif]
Primer efficiency for RT-qPCR analysis.

S1 Table [log2]
List of host mRNAs differentially regulated upon . i infection.

S2 Table [go]
Gene Ontology enriched categories for translationally-regulated mRNAs upon . infection.

S3 Table [cgg]
List of TOP, eIF4A-sensitive, and (CGG) motif-harboring transcripts found in the translationally regulated dataset of . -infected BMDMs.

S4 Table [docx]
Primer sequences used for RT-qPCR analyses.